diff --git a/src/UserGuide/latest/Background-knowledge/Cluster-Concept.md b/src/UserGuide/latest/Background-knowledge/Cluster-Concept.md index d6f57bf2..6b6bb7c8 100644 --- a/src/UserGuide/latest/Background-knowledge/Cluster-Concept.md +++ b/src/UserGuide/latest/Background-knowledge/Cluster-Concept.md @@ -1,32 +1,32 @@ # Cluster-related Concepts + The figure below illustrates a typical IoTDB 3C3D1A cluster deployment mode, comprising 3 ConfigNodes, 3 DataNodes, and 1 AINode: - -This deployment involves several key concepts that users commonly encounter when working with IoTDB clusters, including: -- **Nodes** (ConfigNode, DataNode, AINode); -- **Slots** (SchemaSlot, DataSlot); -- **Regions** (SchemaRegion, DataRegion); +![](https://alioss.timecho.com/docs/img/Common-Concepts_02.png) + +This deployment involves several key concepts that users commonly encounter when working with IoTDB clusters, including: + +- **Nodes** (ConfigNode, DataNode, AINode); +- **Slots** (SchemaSlot, DataSlot); +- **Regions** (SchemaRegion, DataRegion); - **Replica Groups**. The following sections will provide a detailed introduction to these concepts. @@ -34,6 +34,7 @@ The following sections will provide a detailed introduction to these concepts. ## Nodes An IoTDB cluster consists of three types of nodes (processes): **ConfigNode** (the main node), **DataNode**, and **AINode**, as detailed below: + - **ConfigNode:** ConfigNodes store cluster configurations, database metadata, the routing information of time series' schema and data. They also monitor cluster nodes and conduct load balancing. All ConfigNodes maintain full mutual backups, as shown in the figure with ConfigNode-1, ConfigNode-2, and ConfigNode-3. ConfigNodes do not directly handle client read or write requests. Instead, they guide the distribution of time series' schema and data within the cluster using a series of [load balancing algorithms](../Technical-Insider/Cluster-data-partitioning.md). - **DataNode:** DataNodes are responsible for reading and writing time series' schema and data. Each DataNode can accept client read and write requests and provide corresponding services, as illustrated with DataNode-1, DataNode-2, and DataNode-3 in the above figure. When a DataNode receives client requests, it can process them directly or forward them if it has the relevant routing information cached locally. Otherwise, it queries the ConfigNode for routing details and caches the information to improve the efficiency of subsequent requests. - **AINode:** AINodes interact with ConfigNodes and DataNodes to extend IoTDB's capabilities for data intelligence analysis on time series data. They support registering pre-trained machine learning models from external sources and performing time series analysis tasks using simple SQL statements on specified data. This process integrates model creation, management, and inference within the database engine. Currently, the system provides built-in algorithms or self-training models for common time series analysis scenarios, such as forecasting and anomaly detection. @@ -41,19 +42,22 @@ An IoTDB cluster consists of three types of nodes (processes): **ConfigNode** (t ## Slots IoTDB divides time series' schema and data into smaller, more manageable units called **slots**. Slots are logical entities, and in an IoTDB cluster, the **SchemaSlots** and **DataSlots** are defined as follows: + - **SchemaSlot:** A SchemaSlot represents a subset of the time series' schema collection. The total number of SchemaSlots is fixed, with a default value of 1000. IoTDB uses a hashing algorithm to evenly distribute all devices across these SchemaSlots. - **DataSlot:** A DataSlot represents a subset of the time series' data collection. Based on the SchemaSlots, the data for corresponding devices is further divided into DataSlots by a fixed time interval. The default time interval for a DataSlot is 7 days. ## Region In IoTDB, time series' schema and data are replicated across DataNodes to ensure high availability in the cluster. However, replicating data at the slot level can increase management complexity and reduce write throughput. To address this, IoTDB introduces the concept of **Region**, which groups SchemaSlots and DataSlots into **SchemaRegions** and **DataRegions** respectively. Replication is then performed at the Region level. The definitions of SchemaRegion and DataRegion are as follows: -- **SchemaRegion**: A SchemaRegion is the basic unit for storing and replicating time series' schema. All SchemaSlots in a database are evenly distributed across the database's SchemaRegions. SchemaRegions with the same RegionID are replicas of each other. For example, in the figure above, SchemaRegion-1 has three replicas located on DataNode-1, DataNode-2, and DataNode-3. -- **DataRegion**: A DataRegion is the basic unit for storing and replicating time series' data. All DataSlots in a database are evenly distributed across the database's DataRegions. DataRegions with the same RegionID are replicas of each other. For instance, in the figure above, DataRegion-2 has two replicas located on DataNode-1 and DataNode-2. + +- **SchemaRegion**: A SchemaRegion is the basic unit for storing and replicating time series' schema. All SchemaSlots in a database are evenly distributed across the database's SchemaRegions. SchemaRegions with the same RegionID are replicas of each other. For example, in the figure above, SchemaRegion-1 has three replicas located on DataNode-1, DataNode-2, and DataNode-3. +- **DataRegion**: A DataRegion is the basic unit for storing and replicating time series' data. All DataSlots in a database are evenly distributed across the database's DataRegions. DataRegions with the same RegionID are replicas of each other. For instance, in the figure above, DataRegion-2 has two replicas located on DataNode-1 and DataNode-2. ## Replica Groups + Region replicas are critical for the fault tolerance of the cluster. Each Region's replicas are organized into **replica groups**, where the replicas are assigned roles as either **leader** or **follower**, working together to provide read and write services. Recommended replica group configurations under different architectures are as follows: -| Category | Parameter | Single-node Recommended Configuration | Distributed Recommended Configuration | -|:------------:|:-----------------------:|:------------------------------------:|:-------------------------------------:| -| Schema | `schema_replication_factor` | 1 | 3 | -| Data | `data_replication_factor` | 1 | 2 | \ No newline at end of file +| Category | Parameter | Single-node Recommended Configuration | Distributed Recommended Configuration | +| :------: | :-------------------------: | :-----------------------------------: | :-----------------------------------: | +| Schema | `schema_replication_factor` | 1 | 3 | +| Data | `data_replication_factor` | 1 | 2 | diff --git a/src/UserGuide/latest/Background-knowledge/Data-Type.md b/src/UserGuide/latest/Background-knowledge/Data-Type.md index 846e8067..e442c7b5 100644 --- a/src/UserGuide/latest/Background-knowledge/Data-Type.md +++ b/src/UserGuide/latest/Background-knowledge/Data-Type.md @@ -1,22 +1,19 @@ # Data Type @@ -25,40 +22,41 @@ IoTDB supports the following data types: -* BOOLEAN (Boolean) -* INT32 (Integer) -* INT64 (Long Integer) -* FLOAT (Single Precision Floating Point) -* DOUBLE (Double Precision Floating Point) -* TEXT (Long String) -* STRING（String） -* BLOB（Large binary Object） -* TIMESTAMP（Timestamp） -* DATE（Date） - +- BOOLEAN (Boolean) +- INT32 (Integer) +- INT64 (Long Integer) +- FLOAT (Single Precision Floating Point) +- DOUBLE (Double Precision Floating Point) +- TEXT (Long String) +- STRING（String） +- BLOB（Large binary Object） +- TIMESTAMP（Timestamp） +- DATE（Date） + The difference between STRING and TEXT types is that STRING type has more statistical information and can be used to optimize value filtering queries, while TEXT type is suitable for storing long strings. ### Float Precision -The time series of **FLOAT** and **DOUBLE** type can specify (MAX\_POINT\_NUMBER, see [this page](../SQL-Manual/SQL-Manual.md) for more information on how to specify), which is the number of digits after the decimal point of the floating point number, if the encoding method is [RLE](Encoding-and-Compression.md) or [TS\_2DIFF](Encoding-and-Compression.md). If MAX\_POINT\_NUMBER is not specified, the system will use [float\_precision](../Reference/DataNode-Config-Manual.md) in the configuration file `iotdb-system.properties`. +The time series of **FLOAT** and **DOUBLE** type can specify (MAX_POINT_NUMBER, see [this page](../SQL-Manual/SQL-Manual.md) for more information on how to specify), which is the number of digits after the decimal point of the floating point number, if the encoding method is [RLE](Encoding-and-Compression.md) or [TS_2DIFF](Encoding-and-Compression.md). If MAX_POINT_NUMBER is not specified, the system will use [float_precision](../Reference/DataNode-Config-Manual.md) in the configuration file `iotdb-system.properties`. ```sql CREATE TIMESERIES root.vehicle.d0.s0 WITH DATATYPE=FLOAT, ENCODING=RLE, 'MAX_POINT_NUMBER'='2'; ``` -* For Float data value, The data range is (-Integer.MAX_VALUE, Integer.MAX_VALUE), rather than Float.MAX_VALUE, and the max_point_number is 19, caused by the limition of function Math.round(float) in Java. -* For Double data value, The data range is (-Long.MAX_VALUE, Long.MAX_VALUE), rather than Double.MAX_VALUE, and the max_point_number is 19, caused by the limition of function Math.round(double) in Java (Long.MAX_VALUE=9.22E18). +- For Float data value, The data range is (-Integer.MAX_VALUE, Integer.MAX_VALUE), rather than Float.MAX_VALUE, and the max_point_number is 19, caused by the limitation of function Math.round(float) in Java. +- For Double data value, The data range is (-Long.MAX_VALUE, Long.MAX_VALUE), rather than Double.MAX_VALUE, and the max_point_number is 19, caused by the limitation of function Math.round(double) in Java (Long.MAX_VALUE=9.22E18). ### Data Type Compatibility When the written data type is inconsistent with the data type of time-series, + - If the data type of time-series is not compatible with the written data type, the system will give an error message. - If the data type of time-series is compatible with the written data type, the system will automatically convert the data type. The compatibility of each data type is shown in the following table: | Series Data Type | Supported Written Data Types | -|------------------|------------------------------| +| ---------------- | ---------------------------- | | BOOLEAN | BOOLEAN | | INT32 | INT32 | | INT64 | INT32 INT64 | @@ -74,12 +72,10 @@ The timestamp is the time point at which data is produced. It includes absolute Absolute timestamps in IoTDB are divided into two types: LONG and DATETIME (including DATETIME-INPUT and DATETIME-DISPLAY). When a user inputs a timestamp, he can use a LONG type timestamp or a DATETIME-INPUT type timestamp, and the supported formats of the DATETIME-INPUT type timestamp are shown in the table below: -
+::: center **Supported formats of DATETIME-INPUT type timestamp** - - | Format | | :--------------------------: | | yyyy-MM-dd HH:mm:ss | @@ -96,16 +92,14 @@ Absolute timestamps in IoTDB are divided into two types: LONG and DATETIME (incl | yyyy.MM.dd HH:mm:ss.SSSZZ | | ISO8601 standard time format | -
- +::: IoTDB can support LONG types and DATETIME-DISPLAY types when displaying timestamps. The DATETIME-DISPLAY type can support user-defined time formats. The syntax of the custom time format is shown in the table below: -
+::: center **The syntax of the custom time format** - | Symbol | Meaning | Presentation | Examples | | :----: | :-------------------------: | :----------: | :--------------------------------: | | G | era | era | era | @@ -138,25 +132,23 @@ IoTDB can support LONG types and DATETIME-DISPLAY types when displaying timestam | ' | escape for text | delimiter | | | '' | single quote | literal | ' | -
+::: ### Relative timestamp -Relative time refers to the time relative to the server time ```now()``` and ```DATETIME``` time. +Relative time refers to the time relative to the server time `now()` and `DATETIME` time. - Syntax: +Syntax: - ``` - Duration = (Digit+ ('Y'|'MO'|'W'|'D'|'H'|'M'|'S'|'MS'|'US'|'NS'))+ - RelativeTime = (now() | DATETIME) ((+|-) Duration)+ - - ``` +``` +Duration = (Digit+ ('Y'|'MO'|'W'|'D'|'H'|'M'|'S'|'MS'|'US'|'NS'))+ +RelativeTime = (now() | DATETIME) ((+|-) Duration)+ +``` -
+::: center **The syntax of the duration unit** - | Symbol | Meaning | Presentation | Examples | | :----: | :---------: | :----------------------: | :------: | | y | year | 1y=365 days | 1y | @@ -172,13 +164,13 @@ Relative time refers to the time relative to the server time ```now()``` and ``` | us | microsecond | 1us=1000 nanoseconds | 1us | | ns | nanosecond | 1ns=1 nanosecond | 1ns | -
+::: - eg： +eg： - ``` - now() - 1d2h //1 day and 2 hours earlier than the current server time - now() - 1w //1 week earlier than the current server time - ``` +``` +now() - 1d2h //1 day and 2 hours earlier than the current server time +now() - 1w //1 week earlier than the current server time +``` - > Note：There must be spaces on the left and right of '+' and '-'. +> Note：There must be spaces on the left and right of '+' and '-'.